US 12,241,942 B2
Method for estimating aging state of battery and apparatus for performing method therefor
Sukhan Lee, Suwon-si (KR); and A Reum Kim, Suwon-si (KR)
Assigned to Research & Business Foundation Sungkyunkwan University, Suwon-si (KR)
Filed by Research & Business Foundation SUNGKYUNKWAN UNIVERSITY, Suwon-si (KR)
Filed on Dec. 2, 2021, as Appl. No. 17/540,545.
Claims priority of application No. 10-2020-0189903 (KR), filed on Dec. 31, 2020.
Prior Publication US 2022/0206078 A1, Jun. 30, 2022
Int. Cl. G01R 31/392 (2019.01); G01R 31/3842 (2019.01); G06N 3/08 (2023.01)
CPC G01R 31/392 (2019.01) [G01R 31/3842 (2019.01); G06N 3/08 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A method for estimating an aging state of a battery, the method comprising:
receiving measurement values of the battery from a battery management system;
calculating factors representing change characteristics of each of a discharge voltage and a discharge current in each of charge/discharge cycles by using the measurement values of the battery; and
estimating, based on an n-dimensional vector (n is a natural number) including the calculated factors, the aging state of the battery by using a machine learning model that is pre-trained by using, as an input vector, an n-dimensional vector including factors representing change characteristics of each of a discharge voltage and a discharge current in each of charge/discharge cycles of each battery for use in training by using measurement values of the battery for use in training,
wherein the aging state of the battery is estimated while the charge/discharge cycles include at least one cycle in which the battery for use in training is discharged before the battery for training is fully charged and the battery for use in training is charged before the battery for use in training is completely discharged, and a load current pattern of the battery for use in training is varying within each discharge cycle of the charge/discharge cycles, and
wherein the machine learning model is pre-trained while the charge/discharge cycles include at least one cycle in which the battery for use in training is discharged before the battery for training is fully charged and the battery for use in training is charged before the battery for use in training is completely discharged, and a load current pattern of the battery for use in training is varying within each discharge cycle of the charge/discharge cycles.